Original research BMJ Glob Health: first published as 10.1136/bmjgh-2020-002974 on 6 September 2020. Downloaded from Barriers to maternal health services during the Ebola outbreak in three West African countries: a literature review

1,2 3 1 1,3 Piper Yerger, Mohamed Jalloh, Cordelia E M Coltart, Carina King ‍ ‍

To cite: Yerger P, Jalloh M, ABSTRACT Key questions Coltart CEM, et al. Barriers to Introduction The Ebola virus disease (EVD) outbreak in maternal health services during West Africa, affecting Guinea, Liberia and What is already known? the Ebola outbreak in three West from 2014 to 2016, was a substantial public health crisis African countries: a literature ► Prior evidence from the 2014 to 2016 outbreak of with health impacts extending past EVD itself. Access to ► review. BMJ Global Health Ebola virus disease (EVD) in Guinea, Liberia and maternal health services (MHS) was disrupted during the 2020;5:e002974. doi:10.1136/ Sierra Leone depicts a deterioration in access to epidemic, with reductions in antenatal care, facility-­based bmjgh-2020-002974 maternal health services (MHS). deliveries and postnatal care. We aimed to identify and Handling editor Seye Abimbola describe barriers related to the uptake and provision of What are the new findings? MHS during the 2014–2016 EVD outbreak in West Africa. ►► Fear and mistrust were significant barriers to both ►► Additional material is Methods In June 2020, we conducted a scoping review utilisation and provision of MHS. published online only. To view, of peer-­reviewed publications and grey literature from ► Existing healthcare resource scarcity was exacer- please visit the journal online ► relevant stakeholder organisations. Search terms were (http://dx.​ ​doi.org/​ ​10.1136/​ ​ bated during the EVD outbreak. bmjgh-2020-​ ​002974). generated to identify literature that explained underlying ►► Infection control policies impeded provision of access barriers to MHS. Published literature in scientific MHS and reinforced mistrust between patients and journals was first searched and extracted from PubMed healthcare workers (HCWs). Received 26 May 2020 and Web of Science databases for the period between ►► Education to HCWs and communities regarding EVD Revised 22 July 2020 1 January 2014 and 27 June 2020. We hand-­searched was insufficient. Accepted 25 July 2020 relevant stakeholder websites. A ‘snowball’ approach was used to identify relevant sources uncaptured in the What do the new findings imply? systematic search. The identified literature was examined ►► Patient-­provider fear and mistrust should be ad- to synthesise themes using an existing framework. dressed through policy and educational interven- Results Nineteen papers were included, with 26 barriers tions during future EVD outbreaks.

to MHS uptake and provision identified. Three themes ►► Health systems should be equipped to continue rou- http://gh.bmj.com/ emerged: (1) fear and mistrust, (2) health system and tine services during EVD outbreaks. service constraints, and (3) poor communication. Our ►► Prompt education and training to HCWs and com- analysis of the literature indicates that fear, experienced by munities is necessary to facilitate continued access both service users and providers, was the most recurring to MHS. barrier to MHS. Constrained health systems negatively impacted MHS on the supply side. Poor communication and inadequately coordinated training efforts disallowed 2015, the extent to which Millennium Devel- on September 6, 2020 by guest. Protected copyright. competent provision of MHS. opment Goal 5 aimed to reduce maternal 2 © Author(s) (or their Conclusions Barriers to accessing MHS during the EVD morbidity and mortality was broadly missed. employer(s)) 2020. Re-­use outbreak in West Africa were influenced by complex but Continued efforts are essential to improve permitted under CC BY-­NC. No inter-­related factors at the individual, interpersonal, health equitable access to MHS, particularly in the commercial re-­use. See rights and permissions. Published by system and international level. Future responses to EVD context of public health crises. Public health BMJ. outbreaks need to address underlying reasons for fear and crises, such as the 2014 outbreak of Ebola virus 1Institute for Global Health, mistrust between patients and providers, and ensure MHS disease (EVD) in West Africa, risk disrupting University College London, are adequately equipped both routinely and during crises. MHS and progress made in maternal health London, UK 2 outcomes, particularly in regions with weak Care Ring, Children and Family 3 Services Center, Charlotte, North existing health infrastructure. Carolina, USA There is evidence from the 2014 outbreak 3Department of Global Public INTRODUCTION of EVD in Guinea, Liberia and Sierra Leone Health, Karolinska Institutet, Timely access to skilled maternal health that shows a deterioration in uptake and Stockholm, Sweden services (MHS) is a critical component of provision of MHS,4–10 while just two studies, Correspondence to precluding preventable maternal morbidity both conducted in Sierra Leone, reported 1 11 12 Dr Carina King; and mortality. Although maternal mortality insignificant changes in access to MHS. c.​ ​king@ucl.​ ​ac.uk​ decreased 29% globally between 1990 and When compared with preoutbreak metrics,

Yerger P, et al. BMJ Global Health 2020;5:e002974. doi:10.1136/bmjgh-2020-002974 1 BMJ Global Health BMJ Glob Health: first published as 10.1136/bmjgh-2020-002974 on 6 September 2020. Downloaded from significant reductions in facility-based­ deliveries have Inclusion and exclusion been found across the three countries.4–10 Similarly, All literature generated was screened by title, abstract across Guinea, Liberia and Sierra Leone, reported utili- and full text using inclusion and exclusion criteria. We sation of antenatal care (ANC) and postnatal care (PNC) included literature that provided an explanation of the saw significant reductions.5–10 Previous studies have esti- underlying barriers to accessing MHS. Specific inclusion mated that utilisation of the first ANC visit and facility-­ criteria comprised literature in English or French, set in based deliveries in Guinea and PNC in Sierra Leone were Guinea, Liberia and/or Sierra Leone, regarding EVD, found to be reduced by 59%, 62%6 and 22%,13 respec- published between 1 January 2014 and 27 June 2020 and tively. In parts of Liberia, another study reported utilisa- relating to MHS, users or providers. MHS included health tion of ANC and facility-­based deliveries to be at less than services necessary to promote health, manage compli- 14% and 9% of peak utilisation, respectively.5 After the cations and prevent morbidity and mortality related to start of the outbreak, one paper reports a 34% increase pregnancy, birth and the postpartum period, including in the facility-­based maternal mortality ratio in Sierra skilled birth attendance, ANC and PNC.14 Leone13 with other authors estimating similar increases We excluded vaccine research, development and clin- in maternal mortality due to reduced access to MHS.8 ical trials; clinical management, pathological and physio- Prior studies that quantified the negative impact of logical research; case studies and case definitions; chain of EVD on access to MHS did not robustly provide an expla- and outbreak modelling; capacity building, nation for the documented reductions.4–10 With more systems strengthening and strategic planning for future frequent EVD outbreaks occurring since 2014, as seen in outbreaks; assessments of the status of MHS before or the Democratic Republic of Congo, it is essential to quan- after the epidemic; ethics papers; personal commentary; tify negative changes in MHS, and investigate the under- child health, including prevention of mother-to-­ child­ lying causes of the negative changes to inform policies transmission of HIV; press releases and news articles; and and interventions. Against this background, we aimed to reviews, reflections and analyses of existing literature. identify and describe barriers to MHS uptake and provi- sion during the 2014–2016 EVD outbreak in Guinea, Analysis and data extraction Liberia and Sierra Leone to guide recommendations for An existing theoretical framework was applied to guide 15 future EVD outbreak responses. literature analysis. The framework’s authors expanded on existing literature regarding barriers to accessing METHODS healthcare in low-income­ countries.15–17 The following We conducted a scoping literature review using scholarly definition of access was used: and grey literature. The opportunity to reach and obtain appropriate health Search method care services in situations of perceived need for care… re- We searched PubMed and Web of Science in August sulting from the interface between the characteristics of persons, households, social and physical environments and 2019, and updated this in June 2020. Search terms input the characteristics of health systems, organisations and pro- http://gh.bmj.com/ to ‘All Fields’ were: ebola OR ‘ebola virus’ OR ‘ebola viders.18 virus disease’ AND matern* OR obstet* OR perinatal OR antenatal OR pregnan* OR birth* OR ‘child birth’ AND The authors described barriers to access in four cate- health OR ‘health care’ OR service* OR deliver* OR gories: geographic accessibility, availability, afforda- 15 provi* OR utili!e* OR uptake OR system* OR impact* bility and acceptability. Each category is dichotomised OR implication* OR barrier* OR facilit* OR access* OR further into supply side (service provision) and demand 15 on September 6, 2020 by guest. Protected copyright. ‘care seeking’. side (service uptake) barriers. Barriers to service provi- Grey literature not published in scientific journals was sion considered elements at the health system level that included to consider research from key actors in the EVD hampered the ability or desire of providers to deliver 15 response. To facilitate inclusivity, a broad search strategy MHS competently and safely. Barriers to service uptake 15 was employed using the phrase ‘Ebola and maternal hindered users from seeking or receiving MHS. health’ on websites of the following key response actors during the outbreak: WHO, Médecins Sans Frontières, RESULTS Oxfam, United Nations Population Fund (UNFPA), Database searches yielded 454 papers, of which 166 were UNICEF, Centers for Disease Control and Prevention, duplicates. The remaining 288 papers were screened and the Ministries of Health for Liberia, Guinea and and excluded by title (n=237), abstract (n=24) and full Sierra Leone. We performed a hand-­search, or ‘snowball text (n=11), resulting in 16 papers. Hand-searching­ technique’, using the references of literature included, resulted in an additional three papers included (figure 1; and relevant literature reviews, to identify any further online supplemental material 1). Reasons for exclusions sources uncaptured in the systematic search. following abstract review included: papers unrelated to Patient and public involvement MHS (n=5); assessments of MHS before or after the EVD Neither patients nor individuals of the public were outbreak (n=2); capacity building or systems strength- involved in our research at any stage. ening focus (n=2); papers that quantified access to MHS

2 Yerger P, et al. BMJ Global Health 2020;5:e002974. doi:10.1136/bmjgh-2020-002974 BMJ Global Health BMJ Glob Health: first published as 10.1136/bmjgh-2020-002974 on 6 September 2020. Downloaded from

Figure 1 Preferred Reporting Items for Systematic Reviews and Meta-­Analyses (PRISMA) flow diagram. EVD, Ebola virus disease; MHS, maternal health services. without describing underlying barriers (n=13); or papers Theme I: fear and mistrust no longer available (n=2). Exclusions following full-­text Fifteen papers reported fear and mistrust as barriers to http://gh.bmj.com/ reading were reviews (n=3), reflections or analyses (n=7) MHS uptake and/or provision. Three papers made no and one news article. Of the 19 papers included, 15 were mention of fear or mistrust as a barrier to MHS,13 21 22 explicitly peer reviewed and four did not clearly state the while another recognised fear among healthcare workers peer-­review status. The types of literature included were: (HCWs) that did not impede service provision.23 four quantitative studies, five qualitative studies, eight Multidimensional fear cut across every dimension of mixed methods studies, one academic commentary and the healthcare access framework and was experienced by on September 6, 2020 by guest. Protected copyright. one modelling study. Two papers were specific to Guinea, MHS users and providers in Guinea, Liberia and Sierra 4 to Liberia, 11 to Sierra Leone and 2 were general to all Leone. The literature evidenced that pregnant women three countries. A Preferred Reporting Items for System- feared multiple aspects of engaging with healthcare: atic Reviews and Meta-Analyses­ diagram illustrates the acquiring EVD at health facilities; being transferred to results of the literature search (figure 1).19 an EVD treatment centre; the appearance of HCWs in After applying the framework,15 26 barriers to MHS personal protective equipment (PPE); receiving disre- uptake and provision were identified and are presented spectful care; and the possible refusal of MHS.3 24–35 Fear in table 1. and stigmatisation of HCWs was a particularly noteworthy The literature indicates that accessible, available, barrier to MHS given their increased exposure to and affordable and acceptable MHS diminished during the risk of contracting EVD.3 24 26–31 35 Demand-side­ fear had EVD outbreak, with availability and acceptability partic- immediate subsidiary effects as it led to delayed uptake of ularly affected. Patterns in the literature yielded three MHS, which increased acuity of maternal morbidity.24 26 30 themes: fear and mistrust; health system and service Fear was facilitated by existing public mistrust of the constraints; and poor communication. Each theme government and government-backed­ healthcare and represents a synthesis of barriers to MHS for both uptake was exacerbated as rumours regarding EVD and the and provision as reflected in the literature, and particu- intentions of HCWs and other EVD responders prolifer- larly concurs with the research of Bell and colleagues.20 ated.3 24–30 32 35 Authors reported rumours such as HCWs

Yerger P, et al. BMJ Global Health 2020;5:e002974. doi:10.1136/bmjgh-2020-002974 3 BMJ Global Health BMJ Glob Health: first published as 10.1136/bmjgh-2020-002974 on 6 September 2020. Downloaded from

Table 1 Barriers to uptake and provision of maternal health services Provision barriers Uptake barriers Geographic accessibility 35 33 35 ►► Travel restrictions ►► Quarantine and travel restrictions 27 30 ►► Lack of ambulances Availability ►► Insufficient staffing (absenteeism, abandonment, transfer and/or ►► Rumours regarding EVD and the intentions of death related to EVD)21 23 24 27 30–32 36 HCWs and other EVD responders3 24–30 32 35 36 3 30 31 33–35 ►► Unqualified, traditional HCWs 3 22–24 27 28 34 35 ►► Facility closures, reduced hours 23 27 29 32 ►► Waiting time increased with EVD testing 3 20 23 24 26 28 30 31 35 36 ►► Resource diversion and scarcity 23 28 32 ►► EVD screening difficulties 27 31 34 ►► Exclusion of pregnant women from services 13 23 24 26–31 34 35 ►► MHS reduced, suspended, discontinued or unavailable 27 ►► Late or no referral 3 20 24–31 35 36 ►► HCW fear of EVD ►► Absent, insufficient or delayed training on EVD and infection control3 20 24–26 30 31 35 29 ►► Clinical guidelines absent, unclear, impractical ►► Lapse of support to traditional birth attendants and community health workers for MHS referrals35 Affordability 31 ►► Informal fees  ►► Shift from public to private facilities due to fear of EVD in the public sector3 Acceptability 27 ►► Rumours that MHS are no longer free ►► Community fear and/or mistrust of facility-­based 27 3 24–28 30–36 ►► Staff mistrust of pregnant women HCWs and health facilities 31 ►► Worsened interpersonal skills among HCWs ►► Preference for/increased traditional, community-­ based care3 30 31 33–35 ►► Stigmatisation of HCWs, pregnant women and EVD20 24 29 31 33 27 31 35 ►► No touch policy http://gh.bmj.com/ Source: Adapted theoretical framework.15 EVD, Ebola virus disease; HCW, healthcare worker; MHS, maternal health services. injecting EVD27 30 32 or spreading EVD3 35 for money24 25 While the literature suggests that mistrust was a barrier or witchcraft.25 It was further rumoured that HCWs were for women to seek MHS,3 24 25 27 28 30 31 35 those who did on September 6, 2020 by guest. Protected copyright. unnecessarily transferring patients to EVD treatment seek MHS were distrusted by HCWs.27 31 In Sierra Leone, centres for money or other unknown reasons.25 35 One UNFPA reported that HCWs distrusted pregnant women author reported the rumour that Liberians were being to be honest about EVD symptoms or exposure, reducing killed so that westerners could have their kidneys.3 acceptable provision of MHS.27 This fearful attitude Although the details differed, rumours acted as a barrier towards pregnant women was seemingly embodied by 31 to MHS in all three countries. Further, in part as result the community as well. In Liberia, pregnant women of these rumours, one paper concluded that uptake of claimed to be the ‘most feared social group’ after expe- 31 public MHS decreased, while private MHS increased,3 riencing stigmatisation from their communities, which potentially decreasing affordability of MHS. likely reduced demand-­side access to MHS. Across all three countries, reports found that HCWs providing MHS feared acquiring EVD from patients and Theme II: health system and service constraints transmitting it to their families or communities.3 25–31 35 Resource diversion and scarcity This is particularly relevant for MHS due to the height- The literature indicates that health system constraints ened risk of contracting EVD from exposure to bodily posed a barrier to MHS. A lack of resources for MHS fluids that are expelled during .27 This affected such as equipment, drugs, supplies, personnel and, access from the supply side as motivation to provide MHS most critically, PPE, was reported across the three diminished. countries.20 23 26 28 30 31 35 Existing resource scarcity was

4 Yerger P, et al. BMJ Global Health 2020;5:e002974. doi:10.1136/bmjgh-2020-002974 BMJ Global Health BMJ Glob Health: first published as 10.1136/bmjgh-2020-002974 on 6 September 2020. Downloaded from exacerbated during the epidemic as resources were No touch policy diverted away from routine MHS to supplement the EVD As EVD is transmitted via bodily contact with infected response.26 30 35 In Liberia, Bell et al linked lack of PPE to individuals, a strict ‘no touch’ policy was enacted in fear of service provision.20 health systems across Guinea, Liberia and Sierra Leone, which posed a barrier to MHS.27 31 35 It was reported that HCWs in some contexts did not touch patients until Health facilities and personnel 23 27 29 Health facility closure and reduced health facility proven EVD negative The symptoms of EVD often mimic those of obstetric emergencies, and EVD triage hours were reported in Guinea, Liberia and Sierra 29 32 Leone.3 22–24 28 34 35 Three papers, two in Sierra Leone22 23 and screening processes were inefficient. One study in and one in Guinea,24 reported the closure of fewer public Sierra Leone reported that the difficult differential diag- health facilities compared with private health facilities, nosis caused pregnant women to wait in an EVD treat- ment centre until proven EVD negative before they were while just one study conducted in Liberia reported the 29 3 transferred to the maternity unit. This increased waiting inverse. In Sierra Leone, Drevin et al reported that times for MHS, delayed responses to obstetric emergen- closure of private health facilities increased the burden cies and risked mortality for the pregnant woman and of demand for caesarean sections in the public sector.23 her fetus.23 27 29 32 Further, Miller et al reported that HCW Even when facilities remained open, quarantines and confusion regarding the ‘no touch’ policy caused the travel restrictions, imposed to facilitate EVD control, cessation of MHS in some parts of Liberia.35 worsened geographic accessibility to MHS for both service users and some non-governmental­ organisation Theme III: poor communication (NGO)-­staffed providers.33 35 The literature indicates that the EVD response and the Insufficient numbers of HCWs also posed a barrier to ability to continue routine MHS was hampered by poor available MHS in the three countries. HCW shortages communication and coordination at the community, during the outbreak were attributed to job abandon- 26 30 31 35 health system, national and international levels. ment, transfer of HCWs to EVD treatment centres and 21 23 24 27 30 36 As discussed, rumours and misinformation at the commu- HCW death from EVD. A modelling study nity level caused barriers to using MHS, which one paper postulates increases in maternal mortality of 38%, 74% attributed to lack of or delayed community sensitisation and 111% in Guinea, Sierra Leone and Liberia, respec- 35 21 around EVD. At the health system level, lack of knowl- tively, due to EVD-related­ HCW death. Some HCWs edge and absent or delayed training on EVD and ICP abandoned their work due to fear of acquiring EVD from 27 measures were particularly highlighted as barriers to pregnant and labouring women. Among the HCWs that MHS.3 20 24–26 30 31 In Sierra Leone, some midwives received continued service provision, it was reported that some their information about EVD from informal sources, refused care to pregnant women, halting availability of 27 34 rather than reliable national or international govern- MHS in some areas. mental or NGOs, which risked the spread of inaccurate 30 Financial barriers to MHS were addressed in three information and potentially increased fear. Midwives http://gh.bmj.com/ papers. On the demand side, research led by UNFPA also reported a lack of communication on clinical guide- in Sierra Leone reported rumours that previously lines for EVD-­positive pregnant women.29 Even when free health services, including MHS, were no longer evidence-­based information was available and accessible, 27 free during the EVD outbreak. In Liberia, one paper infection control policies and HCW confusion around or reported that informal fees in the public sector, a barrier inability to implement these policies prevented provision to MHS that existed before the EVD outbreak, were exac- of the necessary level of care that MHS require, which on September 6, 2020 by guest. Protected copyright. 31 erbated during the outbreak. On the supply side, it posed supply-­side and demand-­side barriers to MHS.29 was reported in one paper that some community health workers that provided MHS were not paid during the 35 outbreak. DISCUSSION In Liberia and Sierra Leone, the HCW gap in MHS This review explored the barriers to MHS uptake and was partially filled by traditional birth attendants provision during the EVD outbreak in Guinea, Liberia 3 31 33–35 (TBAs). The literature suggests that TBAs were and Sierra Leone by applying an established healthcare 30 33 35 more trusted by pregnant women and less likely access framework. We identified cross-cutting­ themes on 35 to discontinue their provision of MHS due to fear. barriers to MHS uptake and provision. The three recur- However, while the Liberian Ministry of Health recognises ring themes were: fear and mistrust, health system and TBAs, community-based­ births are normally forbidden service constraints, and poor communication. The nega- and referral to facilities is mandatory.3 35 This policy was tive impacts of fear, constrained health systems and poor not enforced during the outbreak, which reduced skilled, communication during the EVD outbreak in West Africa facility-­based MHS.3 35 Because TBAs were not authorised have been discussed by other authors.37–40 Our anal- to attend , they lacked the necessary resources ysis of the literature revealed that fear, experienced by for births and training needed for EVD infection control both service users and providers, was the most recurring and prevention (ICP).35 barrier to MHS across the region. Further, in agreeance

Yerger P, et al. BMJ Global Health 2020;5:e002974. doi:10.1136/bmjgh-2020-002974 5 BMJ Global Health BMJ Glob Health: first published as 10.1136/bmjgh-2020-002974 on 6 September 2020. Downloaded from with Bell et al, our analysis indicates that fear was a driver a successful health system: labour force, health informa- for other barriers to MHS.20 Constrained health system tion, technology and supplies, governance, financing resources negatively impacted MHS provision. Poorly and service provision prior to the outbreak54; this made delivered communication and inadequate training it difficult for local health systems to maintain routine efforts disallowed competent provision of MHS. MHS and effectively respond to the outbreak simultane- Fear played an indelible role in influencing behaviours ously. It is important to bear in mind the broader effects among the public and HCWs during the EVD outbreak of an epidemic and equip routine services for success. In in West Africa.37 41 A consequential effect of fear was the addition to routine MHS, obstetric emergencies require mistrust it propagated between the public and HCWs. prompt facility-­based attention by skilled HCWs, making Moreover, it has been documented that HCWs were consistent access to these services a high priority. Hori- heavily stigmatised partly due to their high incidence zontally strengthening the capacities of national health of EVD infection.40 42–45 HCWs in Guinea, Liberia and systems to provide universal, routine healthcare and the Sierra Leone had 21–32 times more risk of contracting ability to survey and mitigate public health risks necessi- EVD when compared with non-­HCWs.46 Further, HCWs tates continued efforts.55 experienced EVD-related­ mortality rates of 1.45%, 8.07% The authors reported that the EVD outbreak caused a and 6.85% in Guinea, Liberia and Sierra Leone, respec- shift from public, facility-­based MHS to private facilities3 tively.21 Considering the significant risk of morbidity that and TBAs.3 31 33–35 Increased use of private facilities for HCWs encountered provides context to demand-side­ MHS raises questions around equitable access and cost fear of HCWs and health facilities as well as supply-side­ barriers, while the use of TBAs poses a barrier to skilled fear of continuing service provision. MHS. Given that some literature indicated that TBAs A key aspect of the EVD response was to institute strin- were more trusted than facility-­based HCWs,30 33 it may gent protocols for ICP within health facilities, including be prudent to further involve TBAs in future responses the no touch policy and enhancements in the use of to EVD outbreaks by equipping them with the training PPE by HCWs and front-­line responders.45 47–49 These and supplies necessary to mitigate risks of infection and protocols may have introduced some unintended conse- barriers to MHS. quences as the outbreak progressed. For instance, a Mitigating access barriers requires intersectoral collab- qualitative study conducted in Sierra Leone near the oration at the local, national and international levels,15 end of the outbreak found that the PPE worn by health including for effective and efficient responses to public workers facilitated a breakdown in trust between patients health crises. Based on research conducted during the and providers.40 Moreover, some ICP policies were costly SARS epidemic, prioritising quick, widespread dissem- violations of cultural norms, seemed to deepen mistrust ination of health and safety information to HCWs and between communities, patients and HCWs, and further communities is vital.56 Messages should be accessible, facilitated opposition towards government-backed­ simple and culturally appropriate and should rectify healthcare, which subsequently reduced demand-side­ any known misconceptions.56 Further, involving trusted 31 50–52 access to facility-based­ MHS. In other cases where leaders, including TBAs, as the communicators of these http://gh.bmj.com/ mistrust was not a factor, some women may opt out of messages could facilitate dissemination and adherence a facility-based­ delivery for practical reasons under the to the public health recommendations in communities. assumption that HCWs would not touch them. Even as Engaging with and educating communities is essential the outbreak was nearing its end, lingering fear may have to curb transmission of EVD and facilitate confidence also influenced health-­seeking behaviours. For example, and participation in the response.57 Moreover, research a national household survey in Sierra Leone showed that conducted during the outbreak of EVD further indi- on September 6, 2020 by guest. Protected copyright. only 63% of respondents were willing to ride in an ambu- cates that robust, well-coordinated­ education for HCWs lance if they became ill.53 decreased levels of fear40 and may also increase motiva- Lack of PPE for HCWs was particularly highlighted in tion to provide MHS during outbreaks of EVD. Further, the literature as a supply-side­ barrier to MHS. Preout- peer support for HCWs, both in person and via social break resource scarcity may have been less important or media, increased motivation and confidence in providing recognised by HCWs outside times of public health crisis health services.40 as fear of EVD flourished. HCWs must be equipped with Although our review was specific to the 2014 outbreak sufficient resources, particularly PPE, to protect them- of EVD in Guinea, Liberia and Sierra Leone, the lessons selves from infection, mitigate fear and increase provi- learnt may have broader relevance for outbreaks which sion of MHS. Further, it is important to simultaneously incite fear and health system disruptions. Fear of infec- educate the community surrounding the individual and tious diseases certainly preceded EVD56 and health community-­level benefits of PPE to assuage any miscon- systems ill prepared to manage epidemics are not unique ceptions and mistrust that PPE propagates. to Guinea, Liberia and Sierra Leone.58 Indeed, the Health system constraints evidenced in the litera- current COVID-19 pandemic has been projected to result ture are relatively unsurprising given the region’s weak in a significant increase in maternal mortality based on existing infrastructure. Guinea, Liberia and Sierra many of the same barriers.59 Therefore, it is important Leone were deficient in all six of the WHO’s criteria for that historical learning is applied to mitigate the indirect

6 Yerger P, et al. BMJ Global Health 2020;5:e002974. doi:10.1136/bmjgh-2020-002974 BMJ Global Health BMJ Glob Health: first published as 10.1136/bmjgh-2020-002974 on 6 September 2020. Downloaded from impacts of infectious disease outbreaks, and inform the Data availability statement Data sharing not applicable as no data sets were development of resilient health systems. generated and/or analysed for this study. This paper had four key limitations. First, it considers Open access This is an open access article distributed in accordance with the three heterogeneous countries, Guinea, Liberia and Creative Commons Attribution Non Commercial (CC BY-­NC 4.0) license, which permits others to distribute, remix, adapt, build upon this work non-commercially­ , Sierra Leone, together. We identified more literature and license their derivative works on different terms, provided the original work is in Sierra Leone (n=11) and Liberia (n=4) compared properly cited, appropriate credit is given, any changes made indicated, and the with Guinea (n=2). Assessing the region as one entity use is non-commercial.­ See: http://creativecommons.​ ​org/licenses/​ ​by-nc/​ ​4.0/.​ prevented an in-­depth analysis of each country’s social, ORCID iD political and economic situations, which may have posed Carina King http://orcid.​ ​org/0000-​ ​0002-6885-​ ​6716 barriers to MHS. However, themes were found common across the countries suggesting shared experiences and barriers. Second, we focused on access to MHS rather than on quality of MHS. Quality is helpful to consider, REFERENCES as access to poor quality services could increase harm to 1 UNICEF. Maternal and newborn health, 2016. Available: https:// maternal health. Third, the literature could have exag- www.​unicef.​org/​health/​index_​maternalhealth.​html [Accessed 25 Oct 2019]. gerated results, particularly in areas with a high volume 2 Kassebaum NJ, Barber RM, Bhutta ZA, et al. Global, regional, of EVD cases, as the epidemic was an emotionally charged and national levels of maternal mortality, 1990–2015: a systematic analysis for the global burden of disease study 2015. The Lancet situation. Selection bias to ensure safety of researchers 2016;388:1775–812. during the outbreak is plausible, and data collection 3 Gizelis T-­I, Karim S, Østby G, et al. Maternal health care in the time of Ebola: a Mixed-­Method exploration of the impact of the epidemic conducted after its end is subject to recall bias. Further, on delivery services in Monrovia. World Dev 2017;98:169–78. analysis of existing data, such as health facility records, 4 Lori JR, Rominski SD, Perosky JE, et al. A case series study on relies on complete and accurate reporting, which may the effect of Ebola on facility-based­ deliveries in rural Liberia. BMC Pregnancy Childbirth 2015;15:254. have been impossible during the epidemic. Research 5 Iyengar P, Kerber K, Howe CJ, et al. Services for mothers and limitations may create erroneous depictions of barriers newborns during the Ebola outbreak in liberia: the need for improvement in emergencies. PLoS Curr 2015;7. doi:10.1371/ to MHS. Future research that considers more specific currents.outbreaks.4ba318308719ac86fbef91f8e56cb66f. [Epub regional contexts and integrates quality of care into the ahead of print: 16 Apr 2015]. 6 Camara BS, Delamou A, Diro E, et al. Effect of the 2014/2015 Ebola analysis is warranted. Finally, our review did not recog- outbreak on reproductive health services in a rural district of guinea: nise and discuss any resiliency of MHS that was described an ecological study. Trans R Soc Trop Med Hyg 2017;111:22–9. by some authors.11–13 23 7 Delamou A, Ayadi AME, Sidibe S, et al. Effect of Ebola virus disease on maternal and child health services in guinea: a retrospective observational cohort study. Lancet Glob Health 2017;5:e448–57. 8 Sochas L, Channon AA, Nam S. Counting indirect crisis-­related deaths in the context of a low-r­esilience health system: the case of maternal and neonatal health during the Ebola epidemic in Sierra CONCLUSION Leone. Health Policy Plan 2017;32:iii32–9. The 2014 EVD outbreak in West Africa created unique 9 Wagenaar BH, Augusto O, Beste J, et al. The 2014-2015 Ebola virus disease outbreak and primary healthcare delivery in Liberia: time-­ and exacerbated existing barriers to MHS. Patient-­ http://gh.bmj.com/ series analyses for 2010-2016. PLoS Med 2018;15:e1002508. provider fear and mistrust, constrained health systems 10 Shannon FQ, Horace-Kwemi­ E, Najjemba R, et al. Effects of the and insufficient attempts to communicate with HCWs 2014 Ebola outbreak on antenatal care and delivery outcomes in Liberia: a nationwide analysis. Public Health Action 2017;7:88–93. and communities were the key barriers to MHS identi- 11 Quaglio G, Pizzol D, Bome D, et al. Maintaining maternal and child fied. Moreover, fear was seemingly the cause or effect of health services during the Ebola outbreak: experience from Pujehun, Sierra Leone. PLoS Curr 2016;8. doi:10.1371/currents.outbreaks. other barriers to MHS. The global health community, in d67aea257f572201f835772d7f188ba5. [Epub ahead of print: 02 Jun accordance with trusted local leaders, must prioritise stra- 2016]. on September 6, 2020 by guest. Protected copyright. tegic planning to address fear, strengthen health systems 12 Caulker VML, Mishra S, van Griensven J, et al. Life goes on: the resilience of maternal primary care during the Ebola outbreak in rural to continue routine services, and better coordinate Sierra Leone. Public Health Action 2017;7:40–6. communication and education efforts to reduce barriers 13 Jones SA, Gopalakrishnan S, Ameh CA, et al. 'Women and babies are dying but not of Ebola': the effect of the Ebola virus epidemic to MHS in future outbreaks of EVD. on the availability, uptake and outcomes of maternal and newborn health services in Sierra Leone. BMJ Glob Health 2016;1:e000065. Acknowledgements PY would like to thank Christopher O’Brien and Ana 14 WHO. Maternal mortality: key facts, 2018. Available: http://www.​ Rodriguez for providing access to local research facilities. who.​int/​news-​room/​fact-​sheets/​detail/​maternal-​mortality [Accessed Contributors This study was designed by PY and CK. PY conducted the literature 25 Oct 2019]. 15 Jacobs B, Ir P, Bigdeli M, et al. Addressing access barriers to searches, data extraction and analysis. The manuscript was primarily written by health services: an analytical framework for selecting appropriate PY with input from MJ, CEMC and CK. All authors read and approved the submitted interventions in low-­income Asian countries. Health Policy Plan manuscript. 2012;27:288–300. Funding The authors have not declared a specific grant for this research from any 16 Ensor T, Cooper S. Overcoming barriers to health service access: funding agency in the public, commercial or not-­for-­profit sectors. influencing the demand side. Health Policy Plan 2004;19:69–79. 17 Peters DH, Garg A, Bloom G, et al. Poverty and access to health Competing interests None declared. care in developing countries. Ann N Y Acad Sci 2008;1136:161–71. 18 Levesque J-­F, Harris MF, Russell G. Patient-Centr­ ed access to health Patient and public involvement Patients and/or the public were not involved in care: conceptualising access at the interface of health systems and the design, or conduct, or reporting, or dissemination plans of this research. populations. Int J Equity Health 2013;12:18. Patient consent for publication Not required. 19 Moher D, Liberati A, Tetzlaff J, et al. Preferred reporting items for systematic reviews and meta-analyses:­ the PRISMA statement. Provenance and peer review Not commissioned; externally peer reviewed. PLoS Med 2009;6:b2535.

Yerger P, et al. BMJ Global Health 2020;5:e002974. doi:10.1136/bmjgh-2020-002974 7 BMJ Global Health BMJ Glob Health: first published as 10.1136/bmjgh-2020-002974 on 6 September 2020. Downloaded from

20 Bell SA, Munro-­Kramer ML, Eisenberg MC, et al. "Ebola kills 40 Raven J, Wurie H, Witter S. Health workers' experiences of generations": Qualitative discussions with Liberian healthcare coping with the Ebola epidemic in Sierra Leone's health system: a providers. Midwifery 2017;45:44–9. qualitative study. BMC Health Serv Res 2018;18:251. 21 Evans DK, Goldstein M, Popova A. Health-Car­ e worker mortality 41 O'Leary A, Jalloh MF, Neria Y. Fear and culture: contextualising and the legacy of the Ebola epidemic. Lancet Glob Health mental health impact of the 2014-2016 Ebola epidemic in West 2015;3:e439–40. Africa. BMJ Glob Health 2018;3:e000924. 22 Brolin Ribacke KJ, van Duinen AJ, Nordenstedt H, et al. The impact 42 Grinnell M, Dixon MG, Patton M, et al. Ebola Virus Disease in of the West Africa Ebola outbreak on obstetric health care in Sierra Health Care Workers--Guinea, 2014. MMWR Morb Mortal Wkly Rep Leone. PLoS One 2016;11:e0150080. 2015;64:1083–7. 23 Drevin G, Mölsted Alvesson H, van Duinen A, et al. "For this one, let 43 McMahon SA, Ho LS, Brown H, et al. Healthcare providers on the me take the risk": why surgical staff continued to perform caesarean frontlines: a qualitative investigation of the social and emotional sections during the 2014-2016 Ebola epidemic in Sierra Leone. BMJ impact of delivering health services during Sierra Leone's Ebola Glob Health 2019;4:e001361. epidemic. Health Policy Plan 2016;31:1232–9. 24 Barden-O’Fallon­ J, Barry M, Brodish P, et al. Rapid assessment of 44 Sow S, Desclaux A, Taverne B. Ebola in guinea: forms of Ebola-­related implications for reproductive, maternal, newborn and stigmatization of surviving health care workers. Bull Soc Pathol Exot child health service delivery and utilisation in guinea. PLOS Current 2016;109:309–13. Outbreaks 2015;7. 45 CDC. 2014-2016 Ebola outbreak in West Africa, 2019. Available: 25 Dynes MM, Miller L, Sam T, et al. Perceptions of the risk for Ebola https://www.​cdc.​gov/​vhf/​ebola/​history/​2014-​2016-​outbreak/​index.​ and health facility use among health workers and pregnant and html#_​ftn4 [Accessed Mar 2020]. lactating women--Kenema District, Sierra Leone, September 2014. 46 WHO. Health worker Ebola infections in guinea, Liberia, and MMWR Morb Mortal Wkly Rep 2015;63:1226–7. Sierra Leone, 2015. Available: https://​apps.​who.​int/​iris/​bitstream/​ 26 Jones S, Ameh C VSO. Exploring the impact of the Ebola outbreak on routine maternal health services in Sierra Leone, 2015. handle/​10665/​171823/​WHO_​EVD_​SDS_​REPORT_​2015.​1_​eng.​pdf;​ Available: https://www.​vsointernational.​org/​sites/​default/​files/​ jsessionid=​8786​B75E​6ED0​BA3E​CAC2​3730​6D3D95D6?​sequence=1 VSO%​20Sierroa%​20Leone%​20-%​20Impact%​20of%​20Ebola.​pdf [Accessed Mar 2020]. [Accessed 30 Jun 2020]. 47 Frieden TR, Damon IK. Ebola in West Africa--CDC's Role in 27 UNFPA. Rapid assessment of Ebola impact on reproductive health Epidemic Detection, Control, and Prevention. Emerg Infect Dis services and service seeking behaviour in Sierra Leone, 2015. 2015;21:1897–905. Available: https://​reliefweb.​int/​sites/​reliefweb.​int/​files/​resources/​ 48 Hageman JC, Hazim C, Wilson K, et al. Infection Prevention and UNFPA%​20study%​20_​synthesis_​March%​2025_​final.​pdf [Accessed Control for Ebola in Health Care Settings - West Africa and United 30 Jun 2020]. States. MMWR Suppl 2016;65:50–6. 28 Delamou A, Sidibé S, El Ayadi AM, et al. Maternal and child health 49 WHO. Interim infection prevention and control guidance for care of services in the context of the Ebola virus disease: health care patients with suspected or confirmed filovirus haemorrhagic fever workers' knowledge, attitudes and practices in rural guinea. Afr J in health-­care settings, with focus on Ebola, 2014. Available: http:// Reprod Health 2017;21:104–13. www.​euro.​who.​int/__​data/​assets/​pdf_​file/​0005/​268772/​Interim-​ 29 Erland E, Dahl B. Midwives' experiences of caring for pregnant Infection-​Prevention-​and-​Control-​Guidance-​for-​Care-​of-​Patients-​ women admitted to Ebola centres in Sierra Leone. Midwifery with-​Suspected-​or-​Confirmed-​Filovirus-​Haemorrhagic-​Fever-​in-​ 2017;55:23–8. Health-​Care-​Settings,-​with-​Focus-​on-​Ebola-​Eng.​pdf [Accessed Oct 30 Jones S, Sam B, Bull F, et al. 'Even when you are afraid, you stay': 2019]. provision of maternity care during the Ebola virus epidemic: a 50 Wilkinson A, Leach M. Briefing: Ebola-­myths, realities, and structural qualitative study. Midwifery 2017;52:19–26. violence. Afr Aff 2015;114:136–48. 31 Jones T, Ho L, Kun KK, et al. Rebuilding people-­centred maternal 51 Blair RA, Morse BS, Tsai LL. Public health and public trust: survey health services in post-Ebola­ Liberia through participatory action evidence from the Ebola virus disease epidemic in Liberia. Soc Sci research. Glob Public Health 2018;13:1650–69. Med 2017;172:89–97. 32 Black BO. Obstetrics in the time of Ebola: challenges and dilemmas 52 Pellecchia U, Crestani R, Decroo T, et al. Social consequences in providing lifesaving care during a deadly epidemic. BJOG of Ebola containment measures in Liberia. PLoS One 2015;122:284–6. 2015;10:e0143036. 33 Ministry of Social Welfare, Gender and Children’s Affairs, UN 53 Li W, Jalloh MF, Bunnell R, et al. Public Confidence in the Health

Women, Oxfam. Report of the multisector impact assessment of Care System 1 Year After the Start of the Ebola Virus Disease http://gh.bmj.com/ gender dimensions of the Ebola virus disease (EVD) in Sierra Leone, Outbreak - Sierra Leone, July 2015. MMWR Morb Mortal Wkly Rep 2014. Available: https://​awdf.​org/​wp-​content/​uploads/​FINAL-​ 2016;65:538–42. REPORT-​OF-​THE-​Multi-​Sectoral-​GENDER-​Impact-​Assessment_​ 54 Shoman H, Karafillakis E, Rawaf S. The link between the estW Launchedon_​24th-​Feb-​2015_​Family_​kingdom_​Resort.​pdf [Accessed African Ebola outbreak and health systems in guinea, Liberia and 30 Jun 2020]. Sierra Leone: a systematic review. Global Health 2017;13:1. 34 McQuilkin PA, Udhayashankar K, Niescierenko M, et al. Health-­Care 55 WHO. Everybody’s business: strengthening health systems to access during the Ebola virus epidemic in Liberia. Am J Trop Med improve health outcomes: WHO’s framework for action, 2007. Hyg 2017;97:931–6. Available: http://www.​who.​int/​healthsystems/​strategy/​everybodys_​ on September 6, 2020 by guest. Protected copyright. 35 Miller NP, Milsom P, Johnson G, et al. Community health workers during the Ebola outbreak in guinea, Liberia, and Sierra Leone. J business.pdf?​ ​ua=1 [Accessed 25 Oct 2019]. Glob Health 2018;8:020601. 56 Person B, Sy F, Holton K, et al. Fear and stigma: the epidemic within 36 Elston JWT, Moosa AJ, Moses F, et al. Impact of the Ebola outbreak the SARS outbreak. Emerg Infect Dis 2004;10:358–63. on health systems and population health in Sierra Leone. J Public 57 Rugarabamu S, Mboera L, Rweyemamu M, et al. Forty-­Two years of Health 2016;38:673–8. responding to Ebola virus outbreaks in sub-Saharan­ Africa: a review. 37 Shultz JM, Cooper JL, Baingana F, et al. The role of fear-­related BMJ Glob Health 2020;5:e00195:e001955. behaviors in the 2013-2016 West Africa Ebola virus disease 58 Palagyi A, Marais BJ, Abimbola S, et al. Health system preparedness outbreak. Curr Psychiatry Rep 2016;18:104. for emerging infectious diseases: a synthesis of the literature. Glob 38 Strong A, Schwartz DA. Sociocultural aspects of risk to pregnant Public Health 2019;14:1847–68. women during the 2013-2015 multinational Ebola virus outbreak in 59 Roberton T, Carter ED, Chou VB, et al. Early estimates of the indirect West Africa. Health Care Women Int 2016;37:922–42. effects of the COVID-19 pandemic on maternal and child mortality in 39 Elston JWT, Cartwright C, Ndumbi P, et al. The health impact of the low-­income and middle-income­ countries: a modelling study. Lancet 2014-15 Ebola outbreak. Public Health 2017;143:60–70. Glob Health 2020;8:e901–8.

8 Yerger P, et al. BMJ Global Health 2020;5:e002974. doi:10.1136/bmjgh-2020-002974